Another consideration I didn't see mentioned above is the type of wood. Some are quite elastic where others are brittle, and still others are stable while some move drastically. These factors combined with the intended use of the bowl will determine the appropriate thickness. These numerous variables are determined to some extent though practice/experience. If you were turning apple, a thick bowl would likely crack, and a very thin bowl would likely distort drastically (as one obvious example). While a thin bowl of boxelder would be somewhat brittle, but so light its ablity to survive a drop would probably be determined far more by its shape and grain orientation which is yet another variable.
I do have to disagree at least in part with the idea that thin bowls will not survive the "drop test" by virtue of the fact that they are thin. This again is based to some degree on the shape, overall diameter, grain orientation, and type of wood. Vessels made of some woods are drastically lighter and due to their thinness are more flexible, both of which can preserve them when dropped. And although I do alot of thin walled turning, it isn't done just for the sake of turning thin, fine piercing work with an NSK requires 1/16 or thinner with most woods so its a function of the design.
I don't have MM's years of experience, but I have never had a bowl break from dropping it (although I don't honestly make a practice of it). And as I learned from Steve Sinner, a sudden flair in or out at the top of an open vessel can add tremendous strength to a thin object (which can be seen on his pierced cylinders. I have turned lap shades with a slight flair that are 1mm or thinner and they would and have easily survived being dropped.
So the answer to your question is that there are many variables to consider, but you first must start the design based on the intended end use. Whether it is a sculpture or a popcorn bowl, form usually follows function.